X

The new strategy under the Sun

CEO Scott McNealy discusses his company's attempts to outdo Microsoft and IBM, as well as his attitude toward open source and Sun's latest round of layoffs.

Stephen Shankland Former Principal Writer
Stephen Shankland worked at CNET from 1998 to 2024 and wrote about processors, digital photography, AI, quantum computing, computer science, materials science, supercomputers, drones, browsers, 3D printing, USB, and new computing technology in general. He has a soft spot in his heart for standards groups and I/O interfaces. His first big scoop was about radioactive cat poop.
Expertise Processors, semiconductors, web browsers, quantum computing, supercomputers, AI, 3D printing, drones, computer science, physics, programming, materials science, USB, UWB, Android, digital photography, science. Credentials
  • Shankland covered the tech industry for more than 25 years and was a science writer for five years before that. He has deep expertise in microprocessors, digital photography, computer hardware and software, internet standards, web technology, and more.
Stephen Shankland
11 min read
Sun Microsystems CEO Scott McNealy is trying to do some fancy footwork.

His company, a darling during the Internet bubble, has been hit hard by the technology market collapse that followed. On Thursday, Sun announced its third round of layoffs in just under two years.

But McNealy is playing the hand dealt him. He has acquiesced to the popularity of Linux, even while casting doubt on its intellectual property purity. He has accepted Intel's and Advanced Micros Devices' x86 processors into Sun's product line, though chiefly as a strategy to boost the company?s Solaris operating system.

At this week's SunNetwork conference in San Francisco, Sun introduced Java Enterprise System, code-named Orion. It's a collection of server software he hopes will convince customers that it's better to buy integrated collections of hardware and software rather than bits and pieces. It's a gamble, but one that could pay handsome dividends if it works. CNET News.com sat down with McNealy at the close of the conference to get an update on his thinking.

Q: How has Sun's mission changed over the last year or two?
A: The mission is the same as it's been, to solve complex network computing problems.

What's changed in how you're going about it, then?
We are moving through the hierarchy of enlightenment. Call it "McMaslow's hierarchy of IT enlightenment" to self-actualization, where you just use the stuff in the same way you just use telephone switches and you just use nuclear power plants and you use the Hoover Dam, without having to know how to work it.

Click here to Play

McNealy defends value of IT in business
Scott McNealy, CEO, Sun Microsystems
This is why I crack up when I learn my third-grader's learning how to program. I want to go in and tell them, are you teaching him how to program a telephone switch, too? Or work a nuclear power plant? It's just a continuum. We've always done piece parts because people like to buy the piece parts. But now open interfaces, standard building blocks, and providing integratable alternatives to the welded-shut Microsoft hair ball, people are getting more and more comfortable buying less mechanics and more assembled fixtures.

However, with Sun's software strategy, with what was Sun ONE and is now the Java Enterprise System, Sun hasn't been terribly successful in getting that to catch on as piece parts against the other piece parts from competitors. What do you do if your piston rings are not popular, or not good?
You sell them the car. How popular are the BMW piston rings?

Presumably they're good enough.
We've got a way more than good enough J2EE (Java 2 Enterprise Edition) app server. We've got the best directory out there. We've got the best Web server out there. We've got hundreds of millions of people using our mail. We've got the highest-performance Solaris-based file system with the LSC technology that we bought. We've got Sun Cluster as the best clustering environment out there.

You put all this stuff together with the car, and all the sudden we've got a very compelling environment. The piano player comes with the movie, now. If you don't like it, you can still turn the sound off and put your own piano player in there. We don't recommend it, and most people don't do that.

So if the individual parts are really good now, how come people didn't buy them, say, a year ago?
Because we were late with a J2EE app server. We didn't have the services organization to integrate it with all the other pieces. Instead, we hired more engineers to do the integration in the labs. So in some sense, we didn't have first-mover advantage, but we're going to come through and basically change the game on the folks. We don't think directory is an industry; we think it's a feature.

When Sun started, it was the original open-source company (with Berkeley Unix). By trying to embrace open source and evolve what you are doing, are you returning more to your roots?
Open source doesn't matter. It doesn't matter where the bits came from. It matters if it has open interfaces, if it interoperates. You only care: Does it fit in, does it plug in, does it interoperate, does it plug and play?

Actually, the scary part about open source is you don't know where the code came from because nobody will indemnify it. And so we use open source as a technology company to build systems. You saw us build Mad Hatter (Sun's Linux desktop software) almost entirely out of open-source components. You shouldn't try this at home, because you don't have an intellectual property arsenal to fight a SCO (Group) or a Microsoft or somebody else if they came after you, and as a media company you can't ignore copyright. Whereas we can give you software indemnification. We can give you worldwide service and support. And we'll promise you quarterly upgrades of the technology with full integration with the rest of the environment.

Open-sourcing Windows, where Microsoft doesn't give you licenses to the API (application programming interface), will not create multiple sources and will not lower the cost. It is way more important to have open interfaces, even if you have proprietary implementations.

It's been more than a year since Ed Zander (Sun's former president and chief operating officer) left the company.
Who? (Laughs.)

Do you need a separate president and COO?
I believe organizations are much more effective with flatter org charts.

But generally you want somebody who could step in and take over in an emergency.
I've got 13 people reporting to me, any one of which could step in and take over in an emergency. The beauty of it is I've got more to choose from because I've got 13 folks who are there, who are at staff (meetings), who are watching the movie, who know what's going on, who have access to all the information. And we have one level less of indirection. When I say I want to do this, it doesn't go to that person who whispers it to the next person. David Yen (Sun's executive vice president of processors) used to report to (John) Shoemaker who used to report to Zander who used to report to me. Now he reports directly to me. It's a way more balanced way of operating.

But with 13 people who all think they're in charge...
They all have very clear charters.

I'm saying if one of them had to step in...
So we'll have the Alexander Haig thing: "I'm in charge!" (Laughs.)

I've got a very clear statement to my board of what I think should happen if I get run over by a truck. It's up to the board to decide what to do if something happens to me. All that corporate governance and succession planning is very well documented and very well understood and very well laid-out in this company. And by the way, all that succession planning we went through a year-and-a-half ago was very well planned. It was planned a year-and-a-half in advance.

Regarding the most recent batch of targeted layoffs, as opposed to the broader layoffs in October 2001 and October 2002, there are some who criticized Sun for not cutting more deeply earlier, saying if you had cut more then you wouldn't have had to cut as many now. Do you agree?
I don't really understand where they have that perspective because they aren't in there doing the product strategy meetings. They have no idea what our product life cycles are, they have no idea where we are in our IT programs...There are a lot of people out there who are flat-out guessing.

Open source doesn't matter. It doesn't matter where the bits came from.
That doesn't mean we're executing perfectly on getting the exact right people out at the right times. I gave a pretty clear example of where I think we could be using our network operation center to manage and operate our equipment (for customers), yet our customers are still hiring a bunch of people or bringing in IBM Global Services to operate our equipment when we could be doing it electronically, with a bunch of gurus. People will still do the wrong thing for the wrong period of time. If all our customers were to use our network operation center, that would require us to hire a bunch of people, but would allow our customers to get rid of an order or two of magnitude of people in their own organizations. How long will it take us to get there? It'll be awhile.

N1 obviously is an important part of your holistic strategy to sell customers complete systems rather than piece parts. To build N1, you've acquired Terraspring, Pirus Networks and, most recently, CenterRun. Do you now have the intellectual property to build out the full version of N1 or do have to acquire more?
We're throwing other technologies into there. It's all the systems management software that we have on our E15K and 6800 line. What that software does is virtualize and provision the resources inside of the server. Now we're just going to virtualize and provision the stuff outside the server too. We're using Solaris technology, we're using GridWare technology, we're using Jini lookup and discovery technologies. We're using Jxta in the Mad Hatter product for peer-to-peer. All these pieces get built up. Sun Cluster is a whole methodology we'll building into the N1 environment. There are pieces that are missing that we'll invent and/or acquire from the outside.

We'd probably prefer to buy pieces where possible because you take a lot of the engineering risk out and we've got a lot of cash. But there's a lot of times we have to go do it ourselves.

There are a lot of technologies in your vision, like RFID tags...
Cameras, sensors.

You talked today about having a camera in your child's classroom. Are there any privacy drawbacks we should worry about? You said at one point, perhaps somewhat flippantly, "You have no privacy. Get over it."

Absolute anonymity breeds irresponsibility. Audit trails and authentication provide a much more civil society.
But if I have a smart card to authenticate myself to the cable network, and they know who I am, and they know I'm allowed to look at Room 33 in Ormondale School, then only those parents who have kids in the school are allowed to look in. If you have multifactor authentication, like Java card, then there is no privacy issue. I shouldn't be allowed on the grounds unless I have a Java card, I'd feel like my kids were in a safer environment if they had to have multifactor identification to get on the grounds at my kid's school.

Absolute anonymity breeds irresponsibility. Audit trails and authentication provide a much more civil society. I'm just a total believer in that. That's not what the ACLU might argue, but I fear the bad guys more than I fear my government.

What's the most useful task you can do with your $5.7 billion in cash reserves...
Tell everybody on the planet about it.

...besides using it as a marketing ploy to assure people you're going to be around a long time.
Let me tell you why it's important. IBM is no longer under the consent decree where they were not allowed to financially trash-talk their competitors. IBM is running all over the planet saying Sun's not viable and trash-talking us from a financial perspective. I need that war chest. I like our cash position better than IBM's. Our net cash position--take cash minus debt, and pension obligations and see who has a better cash position. You decide. But I think it's very important to be able to respond to the unconstrained IBM trash-talking out there around our financial position.

But are you going to use that?
We are. We bought LSC, Highground, Afara, Pirus, CenterRun, Terraspring, Pixo. We made a lot of wonderful acquisitions that are part of the Lego building blocks that are part of the "Big Frigging Webtone Switch." Only we're trying not to do too many more Cobalts where we use overinflated stock to buy overinflated companies.

Will there come a point where the x86 version of Solaris is going to outship Linux?
I can't make predictions.

The software makers seem to be writing to Linux.
They're writing to Java. And most of the applications written to Linux are running on the app server in Java. You don't write to the operating system anymore because then it's stuck on that OS. You want to write it to the Java level of abstraction because then it runs on the smart card, it runs on the set-top box, it runs on the game machine, it runs on multiple versions of Unix, it runs on multiple versions of Linux, it runs on Mac, it runs on every server running J2EE on the planet. The kernel doesn't matter.

So Oracle, Veritas, BEA Systems, SAP--those companies are writing to the Java standard? I don't think so.
There's the Web services stack, and to me, there's going to be two, probably three long-term. There's the IBM Web services stack, there's the Microsoft Web services stack, and then there's the Solaris-Sun Web services stack, which we just announced, with the most aggressive pricing. It isn't checkmate, maybe, but it's check. We got their king lined up, and right now it's with a pawn. We're not risking a queen going after these things! We're risking very low-revenue part of our business model. I'm not sticking my queen out there to go get nailed out there in the open.

What's your queen? Solaris?
The queen is our systems. Our servers.

You guys have let x86 servers into the fold, so is there a point in having UltraSparc that overlaps?
Because you write to the Web services environment, the kernels don't matter--that's why we can do Linux and Solaris--and the microprocessor instruction set doesn't matter. We can use different microprocessors to run our code. In building our Big Frigging Webtone Switch, there (are) horizontal loads, and there (are) vertical loads, that all have different needs in terms of floating point (math calculations), throughput, watts, retry features, scalability, performance. So what we're doing is growing our chip design investment. David Yen--his R&D is growing faster than any part of the company. But he's picking spots in a more focused way. And we're doing things that Intel isn't doing or AMD isn't doing.

You guys are now saying you'll run your customers' data centers remotely.
We won't run their data centers. We'll operate their equipment. It's a subtle difference. IBM wants to own the data center, wants to own the bricks and mortar. We don't want to own the asset. We just want to operate our equipment for you, remote controlled--and provide the upgrades, do the dynamic reconfiguration, offline a disk drive. And then you can run the data center itself. You can hire (Electronic Data Systems) or (Computer Sciences) to run the data center for you. The last choice would be (Hewlett-Packard) or IBM. Or you can go to a service provider like a Cable and Wireless or an AT&T who'll actually operate the data center for you and connect it to the network.

This is what's intriguing. In the past where you've said you aren't going to compete with your business partners. It seems like you are getting into it now.
No. EDS wants us to go do this. We're not competing. We're a subcontractor to them in this model, and giving them better service and support on our equipment. They have a heck of a time keeping up with our next revs of software, our patches, our new security upgrades. We don't want to own the data centers. We don't want to host these thing--that's the EDS play. Then they do all the software integration and pieces we don't do. They do desktop maintenance. We don't want to do that. We just want to do remote control of the actual piece of equipment. It's very complementary. They're screaming at us to go do it for them.